For example, let's consider treatment of autism by injections of Lupron, a powerful drug used in treatment of advanced prostate cancer and of precocious puberty (development of mature sexual characteristics in preschoolers and children in the early school grades). It is also used essentially as a chemical form of castration for serious sex offenders. Why would any reputable physician use this powerful drug for a purpose for which it has not been approved? Wouldn't a better question be whether one would call reputable a doctor who did this? Let me point out, too, that it is quite possible for a child to be given Lupron without a physician being involved at all. In the Vasquez child abuse case, which I discussed some weeks ago, the adoptive mother asked her pediatrician to give Lupron to a 12-year-old girl, on the grounds that she was maturing too quickly. When the doctor refused, the mother went to Mexico, obtained the drug, and brought it home for administration to the child.

A second example of dangerous "alternative" treatments of autism is the use of chelation therapy, a chemical treatment used in cases of poisoning by heavy metals such as lead. While the dangers of chelation therapy are balanced by its well-documented usefulness in toxic heavy metals, there is no evidence that it is useful for autism-- its harmful side effects are not balanced by a therapeutic effect. Instead, the children receiving this treatment are in potential danger of kidney damage, as well as the loss of minerals needed for good health, such as iron and copper. It is notable that in the past claims were made for the usefulness of chelation therapy in healing diabetic ulcers, reversing gangrene, and slowing the aging process. None of these claims have ever been supported by evidence.

So, what are some better choices for parents of autistic children? What about conventional treatments that appear to be harmless?

Commonly, at least a part of autism treatment is based on Sensory Integration Therapy (SIT). This method, of which a brief glimpse appeared in the "Dateline" program, is based on the idea that autistic children have atypical responses to sensory stimulation and need specific sensory "diets" to help them make sense of the world. SIT may involve brushing of the skin with bristles of different stiffnesses, vestibular stimulation of the kind that results from swings or bouncing on a trampoline, or other types of stimulation. The theory behind SIT has some degree of plausibility, but the research evidence is very mixed and only weakly supports the effectiveness of these techniques. Nevertheless, SIT is popular among occupational therapists, and is often favored by parents, to whom it gives the sense that there is something that they personally can do for their child's condition. SIT is presumably harmless if practiced according to the usual guidelines; my only concern about it is the possibility that parents or practitioners might interpret the use of deep pressure stimulation to mean that physical restraint of the child was appropriate.

The treatment Applied Behavior Analysis (ABA) is a conventional treatment that has been around since the 1970s and is now funded by school systems and other sources. ABA is based on operant conditioning principles and requires intense one-to-one work with a certified therapist for many hours each week. Although ABA advocates have collected and reported much data on outcomes of ABA treatment, the research designs have generally not been at the highest level (see Gernsbacher, M.A. (2003). Is one style of early behavioral treatment for autism "scientifically proven"? Journal of Developmental and Learning Disorders, Vol. 7, pp. 19-25). One study which used random assignment of children to ABA and comparison groups showed some statistically significant improvements in the ABA group, but not as large a proportion of children improving as had been reported in previous, less sophisticated studies. As for potential harm to children, in the very early days of ABA, some aversive treatments, including slapping and electric shock were used, but these have not been part of ABA practice for several decades.

A less-used but very interesting treatment is the Developmental, Individual-difference , Relationship-based approach (DIR), developed some years ago by Stanley Greenspan. This treatment uses social interactions that follow the child's lead and interests to support gradual developmental change, basing the specific activities on a plausible theory of development suggested by Greenspan. The philosophy is like that of "Floortime", an approach to work with young children in the family or in child care settings. Greenspan and his colleagues have reported clinical material from DIR work, but there has been little or no systematic comparison of outcomes from DIR as compared to outcomes from other treatments. It seems quite unlikely that DIR methods could be harmful.

It's discouraging to realize that choices of treatment for autistic children are based less on the existence of good scientific support than on concerns about the potential harm inherent in a practice. That's where we seem to be, however.